CDNPull

CDNPull is a term used to describe a content delivery network (CDN) strategy in which edge servers fetch content from the origin server only when requested by a user and then cache the retrieved copy for subsequent requests. This pull-based approach contrasts with push CDN models where content is proactively pushed to edge caches in advance.

Operation begins when a user requests a resource. The nearest edge cache checks its local store for a valid copy. If the content is cached and fresh, it is served directly. If not, the edge node retrieves the resource from the origin over HTTP(S), optionally validating with cache validators such as ETag or Last-Modified. The retrieved content is then cached at the edge for future requests and served to the user. Cached items are kept until their expiration or eviction according to configured rules.

Key features include adherence to standard HTTP cache controls (such as Cache-Control and max-age), support for varying content and cookies-aware caching, and TTL-based invalidation. CDNPull can be configured to cache static assets, media, or certain dynamic content types while applying rules to govern what is cached and how often it is refreshed. Invalidation and purging mechanisms are used to remove or refresh content as needed to maintain correctness.

Advantages of the pull model include reduced origin bandwidth, improved cache hit ratios as demand patterns emerge, and simplified distribution for large or frequently changing assets. Limitations involve potential latency on the first request for uncached objects, possible stale content if invalidation is delayed, reliance on correct cache headers, and risks during origin outages.

CDNPull is commonly used for media delivery, software updates, and web content where demand is unpredictable or highly variable. See also CDN, edge caching, and cache invalidation.